Crystal forms in pharmaceutical applications: olanzapine, a gift to crystal chemistry that keeps on giving

The rich solid-state chemistry of olanzapine uncovered during the development of this blockbuster drug has inspired the development and application of new tools and techniques for understanding molecular assembly during crystallization.

. DSC-TGA traces of olanzapine monohydrate and dihydrates B, D and E measured at 10 °C/min.
The powder X-ray diffraction (PXRD) pattern of OZPN monohydrate, along with reference patterns for forms I-III and dihydrates B, D, and E, are shown in Figure S2. In being nearly isostructural to forms II and III, the monohydrate PXRD pattern resembles that of the metastable neat polymorphs.
By contrast, the ssNMR spectrum of this hydrate is readily differentiated from the neat polymorphs and hydrates of OZPN, Figure S3. As a superior technique for identifying OZPN forms, solid-state 13 C NMR spectroscopy clearly showed even the best monohydrate material to be contaminated with forms II and III.   Figure S2. PXRD patterns of OZPN forms. Figure S3. Solid-state 13 C NMR spectra of OZPN crystal forms. Form II and III impurity peaks in the monohydrate spectrum are denoted with asterisks (*).  (Reutzel-Edens et al., 2003), was characterized by PXRD and ssNMR spectroscopy for comparison to its partial dehydration product, dihydrate E. The isostructural 2.5 hydrate and dihydrate E show small differences by PXRD, Figure S4. With the incorporation of an additional 0.5 waters of crystallization, the symmetry of the 2.5 hydrate is lower than that of dihydrate E. This is clearly seen by ssNMR spectroscopy, with all of the 13 C peaks of OZPN doubled in the spectrum of the 2.5 hydrate.   As a complement to PIXEL calculations, UNI intermolecular potentials (Gavezzotti, 1994, Gavezzotti & Filippini, 1994 were calculated for OZPN forms I and IV as implemented within Mercury. This method, which uses empirical pair-potential parameters, provides approximate intermolecular interaction energies. As shown in Figure S7, the dispersion bound dimer is the most important pairwise interaction in the crystal structure of form I (UNOGIN01), in agreement with PIXEL calculations (Bhardwaj et al., 2013). The dimer is also stronger than any pairwise interaction in the crystal structure of form IV. Figure S7. UNI intermolecular potentials, showing that the dispersion bound dimer is the most important pairwise interaction in the crystal structure of form I (UNOGIN01) and stronger than any pairwise interaction in catemeric form IV (UNIGON05). Intermolecular interaction energies are given in kJ/mol.

S3. Methods
PXRD patterns were collected on a Bruker D4 Endeaver X-ray powder diffractometer, equipped with a CuK source (=1.54056 Å) and a Vantec detector, and operating at 40 kV and 50 mA, with 0.06 mm divergence and detector slits. Each sample was scanned from 4 to 40º in 0.009º 2 steps at a rate of 0.2 seconds per step.
Cross polarization/magic angle spinning NMR spectra were obtained on a Bruker Avance III 400 wide-bore NMR spectrometer operating at 1 H and 13 C frequencies of 400.131 and 100.622 MHz, respectively, and using either a Bruker 4 mm double-resonance or a 4 mm triple-resonance probe.
The MAS rate was set to 10 kHz using a Bruker MAS-II controller; spinning speeds were maintained within 2 Hz of the set point. SPINAL64 (Fung et al., 2000) decoupling at a proton nutation frequency of 100 kHz was used for heteronuclear decoupling. Spinning sidebands were eliminated by a fivepulse total sideband suppression (TOSS) sequence (Antzutkin, 1999). The CP contact time for transferring magnetization from protons to carbons was set to 1.5 ms and a linear power ramp from was used on the 1 H channel to enhance CP efficiency (Metz et al., 1994). The acquisition time was set to 34 ms and spectra were acquired over a spectral width of 30 kHz with 616 transients and a recycle delay of 5 s. The sample temperature was regulated to 297 ± 1 K in order to minimize frictional heating caused by sample spinning. The 13 C chemical shifts were externally referenced (± 0.05 ppm) to the proton-decoupled 13 C peak of neat (liquid) tetramethylsilane via the high-field resonance of adamantane (δ = 29.5 ppm).
Differential thermal/thermogravimetric analyses were carried out on a TA simultaneous DSC-TGA unit (Model SDT Q600). Samples were heated in open aluminum pans from about 20 to 250 ºC at 10 ºC/min with a nitrogen purge of 100 mL/min. The temperature was calibrated with indium. The weight calibration was performed with a manufacturer-supplied standard.